Phenotypic Plasticity in a Neotropical Treefrog: Predator-Induced Morphologies Vary Due to Evolutionary Background of D. ebraccatus Tadpoles
Annie Innes-Gold ’18, Phoebe Reuben ’17, Hubert Szczygiel ’18, and Justin Touchon (Biology)
Many organisms have variable morphology or coloration which increases crypsis or startles predators, amongst other possible functions. Some organisms are flexible in the sense that they can develop different phenotypes based on varying environmental conditions, a process known as phenotypic plasticity. Tadpoles in particular have become model systems for studying the way that prey animals can respond to cues of predators by developing adaptive morphologies and colors. We measured the phenotypic plasticity of the tadpoles of Dendropsophus ebraccatus, the hourglass or pantless treefrog. This study exposed tadpoles of 43 families of frogs daily to cues of dragonfly larvae (Anax amazili), fish (Astyanax ruberrimus), or predator-free control water. 18 tadpoles were selected from each family and housed individually, six tadpoles being raised in each predation treatment. After 10 days, tadpoles were photographed individually, and morphology and coloration were analyzed using ImageJ. We found that the tadpoles reared with dragonfly cues generally had larger, more colorful tails but smaller overall bodies. However within a treatment, there was substantial variation between families. For example, on average families increased tail redness in response to dragonflies, but the control phenotype for some families had a redder tail then other families could possibly induce, meaning that families do not all have the same starting phenotype. This demonstrates that while all families appear to have the ability to respond plastically to predation risk, there is tremendous genetic variation for the baseline phenotype of different D. ebraccatus families and constraints on the phenotypes that different families can develop. As these frogs have evolved, the selection from each predator has likely been highly variable, and as a result there is no single best phenotype.